Neural Substrates of Language Acquisition · of Language Acquisition ... effortlessly,...

ANRV346-NE31-21 ARI 14 May 2008 12:46

Neural Substratesof Language AcquisitionPatricia Kuhl and Maritza Rivera-GaxiolaInstitute for Learning and Brain Sciences, University of Washington, Seattle,Washington 98195; email: [email protected]

Annu. Rev. Neurosci. 2008. 31:511–34

The Annual Review of Neuroscience is online atneuro.annualreviews.org

This article’s doi:10.1146/annurev.neuro.30.051606.094321

Copyright c! 2008 by Annual Reviews.All rights reserved

0147-006X/08/0721-0511$20.00

Key Wordsbrain measures, language development, speech perception, infants

AbstractInfants learn language(s) with apparent ease, and the tools of modernneuroscience are providing valuable information about the mechanismsthat underlie this capacity. Noninvasive, safe brain technologies havenow been proven feasible for use with children starting at birth. Thepast decade has produced an explosion in neuroscience research exam-ining young children’s processing of language at the phonetic, word, andsentence levels. At all levels of language, the neural signatures of learningcan be documented at remarkably early points in development. Individ-ual continuity in linguistic development from infants’ earliest responsesto phonemes is reflected in infants’ language abilities in the second andthird year of life, a finding with theoretical and clinical implications.Developmental neuroscience studies using language are beginning toanswer questions about the origins of humans’ language faculty.

511

Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

ContentsINTRODUCTION . . . . . . . . . . . . . . . . . . 512WINDOWS TO THE YOUNG

BRAIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513NEURAL SIGNATURES

OF PHONETIC LEARNING. . . . . 515THE ROLE OF SOCIAL FACTORS

IN EARLY LANGUAGELEARNING . . . . . . . . . . . . . . . . . . . . . . 518

NEURAL SIGNATURESOF WORD LEARNING . . . . . . . . . . 520

INFANTS’ EARLY LEXICONS . . . . . . 521NEURAL SIGNATURES OF EARLY

SENTENCE PROCESSING . . . . . . 523BILINGUAL INFANTS: TWO

LANGUAGES, ONE BRAIN . . . . . 526CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 527

INTRODUCTIONLanguage and neuroscience have deep histori-cal connections. Neuroscientists and psycholin-guists have mined the link between languageand the brain since the neuropsychologist PaulBroca described his patient, nicknamed “Tan”for the only word he could utter after his dev-astating accident. “Tan” was found to have abrain lesion in what is now known as Broca’sarea. This finding set the stage for “mapping”language functions in the brain.

Humans’ linguistic capacities have intriguedphilosophers and empiricists for centuries.Young children learn language rapidly andeffortlessly, transitioning from babbling at 6months of age to full sentences by the age of3, following a consistent developmental pathregardless of culture. Linguists, psychologists,and neuroscientists have struggled to explainhow children acquire language and ponder howsuch regularity is achieved if the acquisitionmechanism depends on learning and environ-mental input. Studies of the infant brain usingmodern neuroscience techniques are providingthe first glimpses of the mechanisms underlyingthe human capacity for language and are exam-

ining whether these mechanisms are specific tospeech and to the human species.

Infants begin life with the ability to de-tect differences among all the phonetic distinc-tions used in the world’s languages (Eimas et al.1971, Streeter 1976). Before the end of the firstyear they learn from experience and develop alanguage-specific phonetic capacity and detectnative-language word forms, learning implic-itly and informally (Kuhl et al. 1992, Saffranet al. 1996, Maye et al. 2002, Newport & Aslin2004a). The goal of experiments on infants hasbeen to determine whether the initial state andthe learning mechanisms are speech specific andspecies specific.

Tests on nonhuman animals and in humansusing nonspeech signals have examined thespecies specificity and domain specificity of theinitial state. Animal tests on phonetic percep-tion (Kuhl & Miller 1975) and on the learn-ing strategies for words (Hauser et al. 2002,Newport & Aslin 2004b), as well as tests onhuman infants’ perception of nonspeech sig-nals resembling speech ( Jusczyk et al. 1977),suggested that general perceptual and cognitiveabilities may account for infants’ initial speechperception abilities (Aslin et al. 1998, Saffranet al. 1999). These data promoted a shift intheoretical positions regarding language, onethat advanced the domain-general hypothesisto describe the initial state more strongly thanhad been done previously ( Jusczyk 1997, Kuhl2000, Saffran 2003, Newport & Aslin 2004a,b).

Theoretical attention is increasingly focusedon the mechanisms that underlie languagelearning. Experiments showed that infants’ ini-tial learning from speech experience involvescomputational skills that are not restricted tospeech or to humans (Saffran 2003, Newport &Aslin 2004a). However, there is new evidence tosuggest that social interaction (Kuhl et al. 2003)and an interest in speech (Kuhl et al 2005a) maybe essential in this process. The combination ofcomputational and social abilities may be exclu-sive to humans (Kuhl 2007).

Ultimately, determining the specificity ofhumans’ language capacity will require func-tional measures of the infant brain. Brain

512 Kuhl · Rivera-Gaxiola

Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

measures are ideal for language because infantssimply have to listen—there is no need for themto make an overt response—as new brain tech-nologies record the infant brain’s language pro-cessing. The promise that neural measures willprovide the answer to the initial state query aswell as the query regarding the specificity ofthe learning mechanisms has motivated intensework on the development of brain technologiesthat are safe and feasible with infants and youngchildren.

The goal of this review is to examine theneuroscience techniques now in use withinfants and young children and the data thesetechniques have produced. Neural substratesuncovered using these techniques now ex-tend from the smallest building blocks oflanguage—phonemes—to children’s encodingof early words and to studies examining chil-dren’s processing of the semantic and syntacticinformation in sentences. We review the neuralsignatures of learning at each of these levels oflanguage. Using brain measures, we can nowlink infants’ processing at these various levels oflanguage: Brain measures of perception of theelemental phonetic units in the first year arestrongly associated with infants’ processing ofwords and syntax in the second and third yearof life, showing continuity in the developmentof language. The studies suggest that exposureto language in the first year of life begins to setthe neural architecture in a way that vaults theinfant forward in the acquisition of language.Our goal is to explore the neural mechanismsthat underlie this capacity.

WINDOWS TO THEYOUNG BRAINNoninvasive techniques that examine lan-guage processing in infants and young childrenhave advanced rapidly (Figure 1) and includeelectroencephalography (EEG)/event-relatedpotentials (ERPs), magnetoencephalography(MEG), functional magnetic resonance imag-ing (fMRI), and near-infrared spectroscopy(NIRS).

Phonemes: thecontrasting element inword pairs that signalsa difference inmeaning for a specificlanguage (e.g.,rake-lake; far-fall)

Phonetic units: setsof articulatory gesturesthat constitutephoneme categories ina language

EEG: electroen-cephalography

Event-relatedpotentials (ERPs):scalp recordings thatmeasure electricalactivity in the brainthat can betime-locked to specificsensory stimuli orcognitive processes

Magnetoencephalo-graphy (MEG):measures magneticfields from electricalcurrents produced bythe brain duringsensory, motor, orcognitive tasks

Functional magneticresonance imaging(fMRI): measureschanges in bloodoxygenation levels thatoccur in response toneural firing, allowingprecise localization ofbrain activity

Near-infraredspectroscopy (NIRS):uses infrared light tomeasure changes inblood concentration asan indicator of neuralactivity throughout thecortex

ERPs have been widely used to study speechand language processing in infants and youngchildren (for reviews, see Kuhl 2004, Friederici2005, Conboy et al. 2008). ERPs, part of theEEG, reflect electrical activity that is time-locked to the presentation of a specific sen-sory stimulus (for example, syllables, words) or acognitive process (recognition of a semantic vi-olation within a sentence or phrase). By placingsensors on a child’s scalp, the activity of neu-ral networks firing in a coordinated and syn-chronous fashion in open field configurationscan be measured, and voltage changes occur-ring as a function of cortical neural activity canbe detected. ERPs provide precise time reso-lution (milliseconds), making them well suitedfor studying the high-speed and temporally or-dered structure of human speech. ERP exper-iments can also be carried out in populationswho, because of age or cognitive impairment,cannot provide overt responses. Spatial resolu-tion of the source of brain activation is limited,however.

MEG is another brain-imaging techniquethat tracks activity in the brain with exquisitetemporal resolution. The SQUID (supercon-ducting quantum interference device) sensorslocated within the MEG helmet measure theminute magnetic fields associated with electri-cal currents produced by the brain when it isperforming sensory, motor, or cognitive tasks.MEG allows precise spatial localization of theneural currents responsible for the sources ofthe magnetic fields. Cheour et al. (2004), Kujalaet al. (2004), and Imada et al. (2006) haveshown phonetic discrimination using MEG innewborns and infants in the first year of life.The newest studies employ sophisticated head-tracking software and hardware that allow cor-rection for infants’ head movements and ex-amine multiple brain areas as infants listento speech (Imada et al. 2006). MEG (as wellas EEG) techniques are completely safe andnoiseless.

Magnetic resonance imaging (MRI) can becombined with MEG and/or EEG to pro-vide static structural/anatomical pictures of

www.annualreviews.org • Neural Substrates of Language Acquisition 513

Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

EEG/ERP: Electrical potential changes• Excellent temporal resolution

• Studies cover the life span

• Sensitive to movement

• Noiseless

+5!V

200Fz

–5!V

Inex

pen

sive

NIRS: Hemodynamic changes• Good spatial resolution

• Studies on infants in the first 2 years

• Sensitive to movement

• Noiseless

Mo

der

ate

MEG: Magnetic field changes• Excellent temporal and spatial resolution

• Studies on adults and young children

• Head tracking for movement calibration

• Noiseless

Exp

ensi

ve

fMRI: Hemodynamic changes• Excellent spatial resolution

• Studies on adults and a few on infants

• Extremely sensitive to movement

• Noise protectors needed

Exp

ensi

ve

Standard

Deviant

Vertex

Tragus

Nasion Inion

LH

33 55

1010

44

9988

11

1111

66

22

1212

77

L

Neuroscience techniques used with infants

Figure 1Four neuroscience techniques now used with infants and young children to examine their brain responses to linguistic signals.


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

the brain. Structural MRIs show maturationalanatomical differences in brain regions acrossthe life span. Individual or averaged MRIs canbe used to superimpose the physiological ac-tivity detected by MEG or EEG to refine thespatial localization of brain activities.

fMRI is a popular method of neuroimag-ing in adults because it provides high spatial-resolution maps of neural activity across the en-tire brain (e.g., Gernsbacher & Kaschak 2003).Unlike EEG and MEG, fMRI does not di-rectly detect neural activity, but rather detectsthe changes in blood oxygenation that occurin response to neural activation/firing. Neu-ral events happen in milliseconds; however, theblood-oxygenation changes that they induceare spread out over several seconds, therebyseverely limiting fMRI’s temporal resolution.Few studies have attempted fMRI with chil-dren because the technique requires infants tobe perfectly still and because the MRI deviceproduces loud sounds making it necessary toshield infants’ ears while delivering the soundstimuli (Dehaene-Lambertz et al. 2002, 2006).

Near-infrared spectroscopy (NIRS) alsomeasures cerebral hemodynamic responses inrelation to neural activity (Aslin & Mehler2005). NIRS utilizes near-infrared light tomeasure changes in blood oxy- and deoxy-hemoglobin concentrations in the brain as wellas total blood volume changes in various re-gions of the cerebral cortex. The NIRS systemcan determine the activity in specific regionsof the brain by continuously monitoring bloodhemoglobin level. Reports have begun to ap-pear on infants in the first two years of life,testing infants’ responses to phonemes as well asto longer stretches of speech such as sentences(Pena et al. 2002, Homae et al. 2006, Bortfeldet al. 2007, Taga & Asakawa 2007). As withother hemodynamic techniques such as fMRI,NIRS does not provide good temporal reso-lution. One of the strengths of this techniqueis that coregistration with other testing tech-niques such as EEG and MEG may be possible.

Each of these techniques is being appliedto infants and young children as they listen tospeech, from phonemes to words to sentences.

SQUID:superconductingquantum interferencedevice, a mechanismused to measureextremely weaksignals, such as subtlechanges in the humanbody’s electromagneticenergy field

Magnetic resonanceimaging (MRI):provides staticstructural andanatomicalinformation, such asgrey matter and whitematter, in variousregions of the brain

In the following sections, we review the findingsat each level of language, noting the advancesmade using neural measures.

NEURAL SIGNATURESOF PHONETIC LEARNINGPerception of the phonetic units of speech—thevowels and consonants that constitute words—is one of the most widely studied behaviorsin infancy and adulthood. Phonetic perceptioncan be studied in children at birth and duringdevelopment as they are bathed in a particu-lar language, in adults from different cultures,in children with developmental disabilities, andin nonhuman animals. Phonetic perceptionstudies conducted critical tests of theories oflanguage development and its evolution. Anextensive literature on developmental speechperception exists, and brain measures areadding substantially to our knowledge of pho-netic development and learning (see Kuhl 2004,Werker & Curtin 2005).

Behavioral studies demonstrated that atbirth young infants exhibit a universal capac-ity to detect differences between phonetic con-trasts used in the world’s languages (Eimas et al.1971, Streeter 1976). This capacity is dramat-ically altered by language experience startingas early as 6 months for vowels and by 10months for consonants. Two important changesoccur at the transition in phonetic perception:Native language phonetic abilities significantlyincrease (Cheour et al. 1998; Kuhl et al. 1992,2006; Rivera-Gaxiola et al. 2005b, Sundara et al.2006) while the ability to discriminate phoneticcontrasts not relevant to the language of theculture declines (Werker & Tees 1984, Cheouret al. 1998, Best & McRoberts 2003, Rivera-Gaxiola et al. 2005b, Kuhl et al. 2006).

By the end of the first year, the infant brainis no longer universally prepared for all lan-guages, but instead primed to acquire the lan-guage(s) to which the infant has been exposed.What was once a universal phonetic capacity—phase 1 of development—narrows as learningproceeds in phase 2. This perceptual narrow-ing is widespread, affecting the perception of


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Critical period: atime in development,hypothesized to bedetermined by bothmaturation andlearning during whichlearners are maximallysensitive to experience

Native languageneural commitment(NLNC): learningshapes neural networksto detect the phoneticand word patterns thatdefine a particularlanguage, facilitatinghigher languagelearning

Mismatch negativity(MMN): reflects thebrain’s reaction to achange in an auditorystimulus

signs from American Sign Language (Krentz& Corina 2008, Klarman et al. 2007), as wellas the perception of visual speech informationfrom talking faces in monolingual (Lewkowicz& Ghazanfar 2006) and bilingual infants(Weikum et al. 2007). In all cases, perceptionchanges from a more universal ability to onethat is more restricted and focused on the prop-erties important to the language(s) to which theinfant is exposed. How learning produces thisnarrowing of perceptual abilities has becomethe focus of intense study because it demon-strates the brain’s shaping by experience dur-ing a critical period in language development.Scientific understanding of the critical periodis advancing (see Bongaerts et al. 1995 andKuhl et al. 2005b for review). Data now sug-gest that an adult’s difficulty in learning a sec-ond language is affected not only by maturation( Johnson & Newport 1989, Newport 1990,Neville et al. 1997, Mayberry & Lock 2003),but also by learning itself (Kuhl et al. 2005b).

Kuhl et al. (2005b, Kuhl et al. 2008) ex-plored the relationship between the transi-tion in phonetic perception and later language.This work examined a critical test stemmingfrom the native language neural commitment(NLNC) hypothesis (Kuhl 2004). According toNLNC, initial native language experience pro-duces changes in the neural architecture andconnections that reflect the patterned regular-ities contained in ambient speech. The brain’s“neural commitment” to native-language pat-terns has bidirectional effects: Neural codingfacilitates the detection of more complex lan-guage units (words) that build on initial learn-ing, while simultaneously reducing attention toalternate patterns, such as those of a nonnativelanguage. This formulation suggests that in-fants with excellent native phonetic skills shouldadvance more quickly toward language. Incontrast, excellent nonnative phonetic abilitieswould not promote native-language learning.Nonnative skills reflect the degree to which thebrain remains uncommitted to native-languagepatterns—phase 1 of development—when per-ception is universally good for all phonetic con-trasts. In phase 1, infants are open to all possible

languages, and the brain’s language areas areuncommitted. On this reasoning, infants whoremain in phase 1 for a longer period of time,showing excellent discrimination of nonnativephonetic units, would be expected to show aslower progression toward language.

The NLNC hypothesis received supportfrom both behavioral (Conboy et al. 2005, Kuhlet al. 2005b) and ERP tests on infants (Rivera-Gaxiola et al. 2005a,b; Silven et al. 2006; Kuhlet al. 2008). ERP studies conducted on in-fants (Figure 2a) used both a nonnative con-trast (Mandarin / -t h/ or Spanish /t-d/) anda native contrast /p-t/. The results revealedthat individual variation in both native andnonnative discrimination, measured neurally at7.5 months of age, were significantly correlatedwith later language abilities. As predicted by theNLNC hypothesis, the patterns of predictionwere in opposite directions (Kuhl et al. 2008).The measure used was mismatch negativity,which has been shown in adults to be a neuralcorrelate of phonetic discrimination (Naatanenet al. 1997). The MMN-like ERP componentwas elicited in infants between 250–400 ms forboth contrasts, tested in counterbalanced order(Figure 2b). Better neural discrimination ofthe native phonetic contrast at 7.5 monthspredicted advanced language acquisition at alllevels—word production at 24 months, sen-tence complexity at 24 months, and meanlength of utterance at 30 months of age. In con-trast, greater neural discrimination of the non-native contrast at the same age predicted lessadvanced language development at the same fu-ture points in time—lower word production,less complex sentences, and shorter mean utter-ance length. The behavioral (Kuhl et al. 2005b)and brain (Kuhl et al. 2008) measures, collectedon the same infants, were highly correlated.

The growth of vocabulary clearly showsthis relationship: Better discrimination of thenative contrast resulted in faster vocabularygrowth (Figure 2c, left column), whereas bet-ter discrimination of the nonnative contrast re-sulted in slower vocabulary growth (Figure 2c,right column). Children’s vocabulary growthwas measured from 14 to 30 months and was


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Nonnative

Native

a

b

100 ms

+3 µV

Better600

400

200

0

Poorer

Est

imat

ed n

um

ber

of

wo

rds

pro

du

ced

Native predictor

Age (months)14 18 22 26 30 18 22 26 30

Nonnative predictor

14

c

Figure 2(a) A 7.5-month-old infant wearing an ERP electrocap. (b) Infant ERP waveforms at one sensor location (CZ) for one infant are shownin response to a native (English) and nonnative (Mandarin) phonetic contrast at 7.5 months. The mismatch negativity (MMN) isobtained by subtracting the standard waveform (black) from the deviant waveform (English = red; Mandarin = blue). This infant’sresponse suggests that native-language learning has begun because the MMN negativity in response to the native English contrast isconsiderably stronger than that to the nonnative contrast. (c) Hierarchical linear growth modeling of vocabulary growth between 14and 30 months for MMN values of +1SD and !1SD on the native contrast at 7.5 months (c, left) and vocabulary growth for MMNvalues of +1SD and !1SD on the nonnative contrast at 7.5 months (c, right). Analyses show that both contrasts predict vocabularygrowth but that the effects of better discrimination are reversed for the native and nonnative contrasts. (From Kuhl et al. 2008)

examined using hierarchical linear growthcurve modeling (Raudenbush et al. 2005).Analyses show that both native and nonna-tive discrimination at 7.5 months are signifi-cant predictors of vocabulary growth, but alsothat the effects of good phonetic discrimina-tion are reversed for the native and nonna-tive predictors. A study of Finnish 11-month-old infants replicated this pattern using Finnishand Russian contrasts (Silven et al. 2006).

Rivera-Gaxiola and colleagues (2005a)demonstrated a similar pattern of predictionusing a different nonnative contrast. Theyrecorded auditory ERP complexes in 7- and11-month-old American infants in response toboth Spanish and English voicing contrasts.They found that infants’ responses to the non-native contrast predicted the number of wordsproduced at 18, 22, 25, 27, and 30 months ofage (Rivera-Gaxiola et al. 2005a). Infants show-ing better neural discrimination (larger nega-tivities) to the nonnative contrast at 11 monthsof age produced significantly fewer words at

all ages when compared with infants showingpoorer neural discrimination of the nonnativecontrast. Thus in both Kuhl et al. (2005b, 2008)and Rivera-Gaxiola et al. (2005a), better dis-crimination of a nonnative phonetic contrastwas associated with slower vocabulary devel-opment. Predictive measures now demonstratecontinuity in development from a variety ofearly measures of pre-language skills to mea-sures of later language abilities. Infants’ earlyphonetic perception abilities (Tsao et al. 2004;Kuhl et al. 2005b, 2008; Rivera-Gaxiola et al.2005a), their pattern-detection skills for speech(Newman et al. 2006), and their processing ef-ficiency for words (Fernald et al. 2006) have allbeen linked to advanced later language abili-ties. These studies form bridges between theearly precursors to language in infancy andmeasures of language competencies in earlychildhood that are important to theory build-ing as well as to understanding clinical popula-tions with developmental disabilities involvinglanguage.


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

THE ROLE OF SOCIAL FACTORSIN EARLY LANGUAGE LEARNINGAlthough studies have shown that young in-fants’ computational skills assist language ac-quisition at the phonological and word level(Saffran et al. 1996, Maye et al. 2002), recentdata suggest that another component, social in-teraction, plays a more significant role in earlylanguage learning than previously thought, atleast in natural language-learning situations(Kuhl et al. 2003, Kuhl 2007, Conboy et al.2008). Neuroscience research that reveals thesource of this interaction between social andlinguistic processing will be of future interest.

The impact of social interaction on speechlearning was demonstrated in a study inves-tigating whether infants are capable of pho-netic and word learning at nine months of agefrom natural first-time exposure to a foreignlanguage. Mandarin Chinese was used in thefirst foreign-language intervention experiment(Kuhl et al. 2003). Infants heard 4 native speak-ers of Mandarin (male and female) during 1225-min sessions of book reading and play acrossa 4–6 week period. A control group of infantsalso came into the laboratory for the same num-ber and variety of reading and play sessions butheard only English. Two additional groups wereexposed to the identical Mandarin material overthe same number of sessions via either standardtelevision or audio-only presentation. After ex-posure, Mandarin syllables that are not phone-mic in English were used to test infant learningusing both behavioral (Kuhl et al. 2003) andbrain (Kuhl et al. 2008) tests.

Infants learned from live exposure toMandarin tutors, as shown by comparisons withthe English control group, indicating that pho-netic learning from first-time exposure couldoccur at nine months of age. However, in-fants’ Mandarin discrimination scores after ex-posure to television or audio-only tutors wereno greater than those of the control infants whohad not experienced Mandarin at all (Kuhl et al.2003) (Figure 3). Learning in the live condi-tion was robust and durable. Behavioral tests ofinfant learning were conducted 2–12 days (me-

dian = 6 days) after the final language-exposuresession, and the ERP tests were conducted be-tween 12 and 30 days (median = 15 days) afterthe final exposure session, with no observabledifferences in infant performance as a functionof the delay.

In further experiments, 9-month-oldEnglish-learning infants were exposed toSpanish, and ERPs were used to test learningof both Spanish phonemes and Spanish wordsafter 12 exposure sessions (Conboy & Kuhl2007). The results showed learning of bothSpanish phonemes and words. Moreover, thestudy was designed to test the hypothesisthat social interaction during the exposuresessions would predict the degree to whichindividual infants learned, and this hypothesiswas confirmed. Social factors, such as overallattention during the exposure sessions, andspecific measures of shared visual attentionbetween the infant and tutor predicted thedegree to which individual infants learnedSpanish phonemes and words (Conboy et al.2008).

Social interaction may be essential forlearning in complex natural language-learningsituations—the neurobiological mechanismsunderlying the evolution of language likely uti-lized the kinds of interactional cues availableonly in a social setting (Kuhl 2007). Humansare not the only species in which social interac-tion plays a significant role in communicationlearning. In other species, such as songbirds, so-cial contact can be essential for communicativelearning (see e.g., Immelmann 1969, Baptista &Petrinovich 1986, Brainard & Knudsen 1998,Goldstein et al. 2003).

The importance of social interaction in hu-man language learning is illustrated by its im-pact on children with autism. A lack of inter-est in listening to speech signals is correlatedwith aberrant neural responses to speech andwith the severity of autism symptoms, indicat-ing the tight coupling between language and so-cial interaction (Kuhl et al. 2005a). Typically de-veloping infants prefer speech even from birth(Vouloumanos & Werker 2004), and particu-larly infant-directed (ID) speech (see sidebar on


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

a

b

Television and audio-onlyChinese language intervention

Monolinguallyraised infants

Foreign-language exposure

Mandarin Chinese phonetic discrimination

Television exposureLive exposure

Per

cen

t co

rrec

t

Live Chinese language intervention

American infants exposed to Chinese

American infants exposed to English

75

70

65

60

55

10 – 12

50

45

Televisionexposure

Audio-onlyexposure

10 – 12

Age (months)

Chinese

English

10 – 12

Chance Chance Chance

Figure 3(a) Effects of natural foreign-language intervention using Mandarin Chinese at 9 months of age. Infants experience either live exposureor a television presentation. (b) Live exposure (left panel ) produced significant learning of a Mandarin Chinese phonetic contrast whencompared with a control group exposed only to English, whereas television exposure or audio-alone exposure did not produce learning(middle panel ). Performance after 12 live exposure sessions beginning at 9 months produced performance in American infants (left panel )that equaled that of monolingual Taiwanese infants who had listened to Mandarin for 11 months (right panel ) (from Kuhl et al. 2003).

Motherese) (Fernald & Kuhl 1987, Grieser &Kuhl 1988).

The data on phonetic and word learningfrom a foreign language indicate that at theearliest stages of language learning, social fac-tors play a significant role, perhaps by gatingthe computational mechanisms underlying

language learning (Kuhl 2007). Interactionbetween the brain mechanisms underlyinglinguistic and social processing will be of stronginterest in the future. Neuroscience researchfocused on shared neural systems for percep-tion and action has a long tradition in speech,and interest in “mirror systems” for social


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

MOTHERESE

Adults use a special speech “register,” called “motherese,”whenaddressing infants and children. Motherese has a unique acousticsignature: It is slower, has a higher average pitch, and containsexaggerated pitch contours (Fernald & Simon 1984). Typicallydeveloping infants and children (Fernald & Kuhl 1987), thoughnot children with autism (Kuhl et al. 2005a), prefer Mothereseover adult-directed speech or a nonspeech analog signal whengiven a choice. Motherese exaggerates the phonetic differencesin speech, both for vowels (Kuhl et al. 1997, Burnham et al. 2002)and for consonants (Liu et al. 2007), making the words containedin Motherese easier to discriminate for infants. Some evidenceindicates that Motherese may aid language learners: Motherswho stretch the acoustic cues in phonemes to a greater extentin infant-directed speech early in life have infants who are betterable to hear subtle speech distinctions when tested in the labora-tory months later (Liu et al. 2003). Motherese also facilitates wordrecognition (Thiessen & Saffran 2003). Brain measures of infants’responses to ID speech are enhanced, as shown by NIRS—moreactivation was seen in left temporal areas when infants were pre-sented with ID speech as opposed to backward speech or silence(Pena et al. 2002), and also when six- to nine-month-old infantswere presented with audio-visual ID speech (Bortfeld et al. 2007).

cognition (Kuhl & Meltzoff 1996, Meltzoff& Decety 2003, Rizzolatti & Craighero 2004,Pulvermuller 2005, Rizzolatti 2005, Kuhl2007) has reinvigorated this tradition (seesidebar on Mirror Systems and Speech). Neu-roscience studies using speech and imagingtechniques have the capacity to examine thisquestion in infants from birth (e.g., Imada et al.2006).

NEURAL SIGNATURESOF WORD LEARNINGA sudden increase in vocabulary typically occursbetween 18 and 24 months of age—a “vocabu-lary explosion” (Ganger & Brent 2004, Fernaldet al. 2006)—but word learning begins muchearlier. Infants show recognition of their ownname at four and a half months (Mandel et al.1995). At six months, infants recognize theirown names or the word Mommy as a word seg-

mentation cue (Bortfeld et al. 2005) and lookappropriately to pictures of their mothers orfathers when hearing “Mommy” or “Daddy”(Tincoff & Jusczyk 1999). By 7 months, in-fants listen longer to passages containing wordsthey previously heard rather than passages con-taining words they have not heard ( Jusczyk &Hohne 1997), and by 11 months infants preferto listen to words that are highly frequent inlanguage input over infrequent words (Halle &de Boysson-Bardies 1994).

One question has been how infants recog-nize potential words in running speech giventhat speech is continuous and has no acousticbreaks between words (unlike the spaces be-tween words in written text). Behavioral stud-ies indicate that by 8 months, infants use var-ious strategies to identify potential words inspeech. For example, infants treated adjacentsyllables with higher transitional probabilitiesas word-like units in strings of nonsense sylla-bles (Saffran et al. 1996, Saffran 2003, Newport& Aslin 2004a); real words contain higher tran-sitional probabilities between syllables and thisstrategy would provide infants with a reliablecue. Infants also use the typical pattern of stressin their language to detect potential words.English, for example, puts emphasis on the firstsyllable, as in the word BASEball, whereas otherlanguages, such as Polish, emphasize the sec-ond). English-learning infants treated syllablescontaining stress as the beginning of a poten-tial word (Cutler & Norris 1988, Johnson &Jusczyk 2001, Nazzi et al. 2006, Hohle et al.2008). Both transitional probabilities betweenadjacent syllables and stress cues thus provideinfants with clues that allow them to identifypotential words in speech.

How is early word recognition evidenced inthe brain? ERPs in response to words indexword familiarity as early as 9 months of ageand word meaning by 13–17 months of age:ERP studies have shown differences in ampli-tude and scalp distributions for components re-lated to words that are known to the child vs.those that are unknown to the child (Molfese1990; Molfese et al. 1990, 1993; Mills et al.1993, 1997, 2005; Thierry et al. 2003).


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Toddlers with larger vocabularies showeda larger N200 for known words vs. unknownwords at left temporal and parietal electrodesites. In contrast, children with smaller vo-cabularies showed brain activation that wasmore broadly distributed (Mills et al. 1993).This distributional difference, with more focal-ized brain activation linked to greater vocabu-lary skill, also distinguished typically develop-ing preschool children from preschool childrenwith autism (Coffey-Corina et al. 2007). In-creased focalization of brain activation has alsobeen seen in adults for native- as opposed tononnative phonemes and words in MEG studies(Zhang et al. 2005). This indicates that focal ac-tivation has the potential to index language ex-perience and proficiency not just in childhood,but over the life span.

Mills et al. (2005) used ERPs in 20-month-old toddlers to examine new word learning. Thechildren listened to known and unknown wordsand to nonwords that were phonotactically legalin English. ERPs were recorded as the childrenwere presented with novel objects paired withthe nonwords. After the learning period, ERPsto the nonwords that had been paired with novelobjects were shown to be similar to those ofpreviously known words, suggesting that newwords may be encoded in the same neural re-gions as previously learned words.

ERP studies on German infants reveal thedevelopment of word-segmentation strategiesbased on the typical stress patterns of Germanwords. When presented with bisyllabic stringswith either stress on the first syllable (typi-cal in German) or the second syllable, infantswho heard first-syllable stress patterns embed-ded in a string of syllables showed the N200ERP component similar to that elicited in re-sponse to a known word, whereas infants pre-sented with the nonnative stress pattern showedno response (Weber et al. 2004). The data sug-gest that German infants at this age are applyinglearned stress rules about their native languageto segment nonsense speech strings into word-like units, in agreement with behavioral data(Hohle et al. In press).

MIRROR SYSTEMS AND SPEECH

The field has a long tradition of theoretical linkage between per-ception and action in speech. The Motor Theory (Liberman et al.1967) and Direct Realism (Fowler 1986) posited close interac-tion between speech perception and production. The discov-ery of mirror neurons in monkeys that react both to the sightof others’ actions and to the same actions they themselves pro-duced (Rizzolatti et al. 1996, 2002; Gallese 2003) has rekindledinterest in a potential mirror system for speech, as has workon the origins of infant imitation (Meltzoff & Moore 1997).Liberman & Mattingly (1985) view the perception-action linkfor speech as potentially innate, whereas Kuhl & Meltzoff (1982,1996) view it as forged early in development through experi-ence. Two new infant studies shed light on the developmentalissue. Imada et al. (2006) used magnetoencephalography (MEG)to study newborns, 6-month-old infants, and 12-month-old in-fants while they listened to nonspeech, harmonics, and syllables(Figure 4). Dehaene-Lambertz et al. (2006) used fMRI to scanthree-month-old infants while they listened to sentences. Bothstudies showed activation in brain areas responsible for speechproduction (the inferior frontal, Broca’s area) in response to au-ditorally presented speech. Imada et al. reported synchronizedactivation in response to speech in auditory and motor areas at 6and 12 months, and Dehaene-Lambertz et al. reported activationin motor speech areas in response to sentences in three-month-olds. Is activation of Broca’s area to the pure perception of speechpresent at birth? Newborns tested by Imada et al. showed no acti-vation in motor speech areas for any signals, whereas auditory ar-eas responded robustly to all signals, suggesting that perception-action linkages for speech develop by three months of age asinfants produce vowel-like sounds. Further work must be doneto determine whether binding of perception and action requiresexperience. Using the tools of modern neuroscience, we can nowask how the brain systems responsible for speech perception andproduction forge links in early development, a necessary step forcommunication development.

INFANTS’ EARLY LEXICONSYoung children’s growing lexicons must codewords in a way that distinguishes them fromone another. Words—produced by differentspeakers, at different rates, and in differentcontexts—have variations that are not relevantto word meaning. Infants have to code the


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Auditory (superior temporal) Wernicke’s areaNewborns 6-month-olds 12-month-olds

Motor (inferior frontal) Broca’s areaNewborns 6-month-olds 12-month-olds

1 2 3 4 5 6

Z score relative to 100 to 0 ms baseline

Figure 4(top) Neuromagnetic signals were recorded in newborns, 6-month-old (shown)and 12-month-old infants in the MEG machine while listening to speech(shown) and nonspeech auditory signals. (bottom) Brain activation in response tospeech recorded in auditory (top row) and motor (bottom row) brain regionsshowed no activation in the motor speech areas in the newborn in response toauditory speech, but increasing activity that was temporally synchronizedbetween the auditory and motor brain regions was seen in 6- and 12-month-oldinfants when listening to speech (from Imada et al. 2006).

critical features of words in sufficient detail toallow them to be distinguished.

One approach to this question is totest how children of different ages re-act to mispronounced words. Reactions tomispronunciations—examining whether chil-dren accept “tup” for cup or “bog” for dog—provide information about the level of phono-logical detail in their mental representations ofwords.

Studies across languages showed that byone year of age infants do not accept mis-pronunciations of common words ( Jusczyk &Aslin 1995, Fennell & Werker 2003), wordsin stressed syllables (Vihman et al. 2004), ormonosyllabic words (Swingley 2005), indicat-ing that their representations of these wordsare well-specified by that age. Other studies us-ing visual fixation of two targets (e.g., apple andball) while one is named (“Where’s the ball?”)indicated that between 14 and 25 months ofage children’s tendencies to fixate the targetitem when it is mispronounced diminishes overtime (Swingley & Aslin 2000, 2002; Bailey &Plunkett 2002, Ballem & Plunkett 2005).

Studies also indicate that when learningnew words, 14-month-old children’s phonolog-ical skills are taxed. Stager & Werker (1997)demonstrated that 14-month-old infants failedto learn new words when similar-sounding pho-netic units were used to distinguish those words(“bih” and “dih”) but did learn if the two newwords were distinct phonologically (“leef” and“neem”). By 17 months of age, infants learnedto associate similar-sounding nonsense wordswith novel objects (Bailey & Plunkett 2002,Werker et al. 2002). Infants with larger vocabu-laries succeeded on this task even at the youngerage, suggesting that infants with greater pho-netic learning skills acquire new words morerapidly. This is consistent with studies showingthat children with better native-language pho-netic learning skills showed advanced vocabu-lary development (Tsao et al. 2004; Kuhl et al.2005b, 2008; Rivera-Gaxiola et al. 2005a).

ERP methods corroborated these results.Mills et al. (2004) compared children’s ERP re-sponses when responding to familiar words thatwere either correctly pronounced or mispro-nounced, as well as some nonwords (Figure 5).At the earliest age tested, 14 months, a negativeERP component (N200–400) distinguishedknown vs. dissimilar nonsense words (“bear” vs.“kobe”) but not known vs. phonetically sim-ilar nonsense words (“bear” vs. “gare”). By20 months, this same ERP component distin-guished correct pronunciations, mispronuncia-tions, and nonwords, supporting the idea that


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

between 14 and 20 months, children’s phono-logical representations of early words becomeincreasingly detailed.

Other evidence of early processing limita-tions stems from infants’ failure to learn a novelword when its auditory label closely resemblesa word they already know (“gall,” which closelyresembles “ball”), suggesting lexical competi-tion effects (Swingley & Aslin 2007). How pho-netic and word learning interact—and whetherthe progression is from phonemes to words,words to phonemes, or bi-directional—is atopic of strong interest that will be aided byusing neuroscientific methods.

Two recent models/frameworks of early lan-guage acquisition, the native language mag-net theory, expanded (NLM-e) (Kuhl et al.2008) and the processing rich informationfrom multidimensional interactive representa-tions (PRIMIR) framework (Werker & Curtin2005), suggest that phonological and wordlearning bidirectionally influence one another.According to NLM-e, infants with better pho-netic learning skills advance more quickly to-ward language because phonetic skills assistphonotactic pattern and word learning (Kuhlet al. 2005b, 2008). At the same time, the morewords children learn, the more crowded lexicalspace becomes, pressuring children to attendto the phonetic units that distinguish wordsfrom one another (see Swingley & Aslin 2007for discussion). Further studies examining bothphoneme and word learning in the same chil-dren will help address this issue.

NEURAL SIGNATURES OF EARLYSENTENCE PROCESSINGTo understand sentences, a child must haveexquisite phonological abilities that allow seg-mentation of the speech signal into words, andthe ability to extract word meaning. In addi-tion, the relationship among words composingthe sentence—between a subject, its verb, andits accompanying object—must be decipheredto arrive at a full understanding of the sen-tence. Human language is based on the ability

Known

Phonetically similar

Phonetically dissimilar

–6

–3

–4

–5

–2

–1

0L

oo

kin

g s

core

Age (months)

N200–400 mean area

14-month-olds 20-month-olds

Figure 5ERP responses to known words, phonetically similar nonsense words, andphonetically distinct nonsense words at 14- and 20-months of age. At14 months, infants’ brain responses are indistinguishable for known andsimilar-sounding nonsense words, although they were distinct for dissimilarnonsense words. By 20 months, infants’ brain responses are distinct for all threetypes of words (from Mills et al. 2005).

Native languagemagnet model,expanded (NLM-e):proposes that earlylanguage experienceshapes neuralarchitecture, affectinglater languagelearning, and bothcomputational andsocial abilities affectlearning

Processing richinformation frommultidimensionalinteractiverepresentations(PRIMIR): aframework for earlylanguage that linkslevels of processingand describes theirinteraction

to process hierarchically structured sequences(Friederici et al. 2006).

Electrophysiological components of sen-tence processing have been recorded in chil-dren and contribute to our knowledge of whenand how the young brain decodes syntactic andsemantic information in sentences. In adults,specific neural systems process semantic vs. syn-tactic information within sentences, and theERP components elicited in response to syn-tactic and semantic anomalies are well estab-lished (Figure 6). For example, a negative ERPwave occurring between 250 and 500 ms thatpeaks around 400 ms, referred to as the N400,is elicited to semantically anomalous words insentences (Kutas 1997). A late positive wavepeaking at "600 ms and that is largest at parietalsites, known as the P600, is elicited in responseto syntactically anomalous words in sentences(Friederici 2002). And a negative wave overfrontal sites between 300 and 500 ms, known


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

LAN

N400

P600

No violation: “The woman was feeding her baby”

Semantic violation: “The woman was feeding her picture”

Syntactic violation: “The woman was feed her baby”

Anterior

Posterior200 ms

Figure 6Distribution and polarity of the ERP responses to normal sentences and sentences with either semantic or syntactic anomalies in adults(from Kuhl & Damasio 2008).

Phonotacticpatterns: sequentialconstraints governingpermissible strings ofphonemes in a givenlanguage. Eachlanguage allowsdifferent sequences

as the late anterior negativity (LAN), is elicitedin response to syntactic and morphological vi-olations (Friederici 2002).

Beginning in a child’s second year of life,ERP data on sentence processing in childrensuggest that adult-like components can beelicited in response to violations in semanticand syntactic components, but that differencesexist in the latencies and scalp distributions ofthese components in children vs. those in adults(Harris 2001; Oberecker et al. 2005; Friedrich& Friederici 2005, 2006; Silva-Pereyra et al.2005a,b, 2007; Oberecker & Friederici 2006).Holcomb et al. (1992) reported the N400 in re-sponse to semantic anomaly in children fromfive years of age to adolescence; the latency ofthe response declined systematically with age(see also Neville et al. 1993, Hahne et al. 2004).Studies also show that syntactically anomaloussentences elicit the P600 in children between 7and 13 years of age (Hahne et al. 2004).

Recent studies have examined these ERPcomponents in preschool children. Harris

(2001) reported an N400-like effect in 36–38-month-old children; the N400 was largestover posterior regions of both hemispheres.Friedrich & Friederici (2005) observed anN400-like wave to semantic anomalies in 19-and 24-month-old German-speaking children.

Silva-Pereyra et al. (2005b) recorded ERPsin children between 36 and 48 months of agein response to semantic (“My uncle will blowthe movie”) and syntactic anomalies (“My unclewill watching the movie”) when compared withcontrol sentences. In both cases, the ERP ef-fects in children were more broadly distributedand elicited at later latencies than in adults.In work with even younger infants (30-month-olds), Silva-Pereyra et al. (2005a) used the samestimuli and observed late positivities distributedbroadly posteriorally in response to syntac-tic anomalies and anterior negativities in re-sponse to semantically anomalous sentences.In both cases, children’s latencies were longerthan those seen in older children and adults(Figure 7), a pattern seen repeatedly and


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Fz

Cz

N

P

Pz

Fz

Cz

Pz

c

a b

100

0

4000 800 1200

–5

µV

%

Time (ms)

0

Figure 7ERP waveforms elicited from 36- and 48-month-old children in response to sentences with syntactic (a) orsemantic (b) violations. Children’s ERP responses resemble those of adults (see Figure 6) but have longerlatencies and are more broadly distributed (c) (from Silva-Pereyra et al. 2005b).


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

attributed to the immaturities of the developingneural mechanisms.

Syntactic processing of sentences withsemantic content information removed—“jabberwocky sentences”—has also been testedusing ERP measures with children. Silva-Pereyra and colleagues (2007) recorded ERPsto phrase structure violations in 36-month-oldchildren using sentences in which the contentwords were replaced with pseudowords whilegrammatical function words were left intact.The pseudowords differed from real wordsby only a few phonemes: “My uncle watcheda movie about my family” (control sentence)became “My macle platched a flovie aboutmy garily” (jabberwocky sentence withoutviolation) and “My macle platched abouta flovie my garily” (jabberwocky sentencewith phrase structure violation). The ERPcomponents elicited to the jabberwocky phrasestructure violations differed as compared withthe same violations in real sentences. Twonegative components were observed: one from750–900 ms and the other from 950–1050 ms,rather than the positivities seen in response tophrase structure violations in real sentences inthe same children. Jabberwocky studies withadults (Munte et al. 1997, Canseco-Gonzalez2000, Hahne & Jescheniak 2001) have alsoreported negative-going waves for jabberwockysentences, though at much shorter latencies.

Recent studies on language and pre-readingskills in 5-year old children prior to their en-try into school (Raizada et al. 2008), as well asin first- through third-grade children (Nobleet al. 2006, 2007), indicate that in healthy, socio-economically diverse populations assessed withvarious linguistic, cognitive, and environmentalmeasures, significant correlations exist betweenlow socio-economic status and behavioral andbrain (fMRI) measures of language skills. Fu-ture work on the complex socio-economic fac-tors that potentially mediate language and read-ing skills in children could potentially leadto interventions that would improve the skillsneeded for reading prior to the time childrenenter school.

BILINGUAL INFANTS: TWOLANGUAGES, ONE BRAINOne of the most interesting questions is howinfants map two distinct languages in the brain.From phonemes to words, and then to sen-tences, how do infants simultaneously bathedin two languages develop the neural networksnecessary to respond in a native-like manner totwo different codes?

Word development in bilingual childrenhas just begun to be studied using ERP tech-niques. Conboy & Mills (2006) recorded ERPsto known and unknown English and Spanishwords in bilingual children at 19–22 months.Expressive vocabulary sizes were obtained inboth English and Spanish and used to deter-mine language dominance for each child. Aconceptual vocabulary score was calculated bysumming the total number of words in bothlanguages and then subtracting the number oftimes a pair of conceptually equivalent words(e.g., “water” and “agua”) occurred in the twolanguages.

ERP differences to known and unknownwords in the dominant language occurred asearly as 200–400 and 400–600 ms in these 19-to 22-month-old infants and were broadly dis-tributed over the left and right hemispheres,resembling patterns observed in younger (13-to 17-month-old) monolingual children (Millset al. 1997). In the nondominant language ofthe same children, these differences were notapparent until late in the waveform, from 600to 900 ms. Moreover, children with high vs.low conceptual vocabulary scores produced agreater number of responses to known words inthe left hemisphere, particularly for the domi-nant language (Conboy & Mills 2006).

Using ERPs, Neville and colleagues haveshown that longer latencies and/or durationsof semantic ERP effects are noted for later-vs. early-acquired second languages across in-dividuals (Neville et al. 1992, Weber-Fox &Neville 1996, Neville et al. 1997). In the syn-tactic domain, ERP effects typically elicited bysyntactically anomalous sentences and closed-vs. open-class words have been absent or


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

attenuated for second languages acquired afterinfancy (Neville et al. 1992, 1997; Weber-Fox& Neville 1996).

A second question about bilingual languagedevelopment is whether it takes longer to ac-quire two languages as compared with acquir-ing one. Bilingual language experience couldimpact the rate of language development, notbecause the learning process is different butbecause learning might require more time forsufficient data from both languages to be ex-perienced. Infants learning two languages si-multaneously might therefore reach the tran-sition from universal phonetic listening (phase1 of development) to language-specific listen-ing (phase 2 in development) at a later pointin development than do infants learning eitherlanguage monolingually. This delay could de-pend on factors such as the number of peoplein the infants’ environment producing the twolanguages in speech directed toward the childand the amount of input these people provide.Such factors could change the development ratein bilingual infants.

Very little data address this question thusfar, and what data do exist are mixed. Somestudies suggest that infants exposed to two lan-guages show a different pattern of phoneticperception development when compared withmonolingual infants (Bosch & Sebastian-Galles2003a,b). Other studies report that phoneticperception in bilingual infants is identical tothat occuring in monolingual infants (Burnset al. 2007). The data are at present not suf-ficient to allow us to answer how bilingual vs.monolingual language exposure affects pho-netic perception development and the timingof the transition in phonetic perception fromone of universal phonetic perception to one re-flecting the language(s) to which infants havebeen exposed.

CONCLUSIONSKnowledge of infant language acquisition isnow beginning to reap benefits from informa-tion obtained by experiments that directly ex-amine the human brain’s response to linguis-

tic material as a function of experience. EEG,MEG, fMRI, and NIRS technologies—all safe,noninvasive, and proven feasible—are now be-ing used in studies with very young infants, in-cluding newborns, as they listen to the phoneticunits, words, and sentences of a specific lan-guage. Brain measures now document the neu-ral signatures of learning as early as 7 monthsfor native-language phonemes, 9 months for fa-miliar words, and 30 months for semantic andsyntactic anomalies in sentences.

Studies now show continuity from the earli-est phases of language learning in infancy tothe complex processing evidenced at the ageof three when all typically developing childrenshow the ability to carry on a sophisticatedconversation. Individual variation in language-specific processing at the phonetic level—at thecusp of the transition from phase 1, in which allphonetic contrasts are discriminated, to phase2, in which infants focus on the distinctionsrelevant to their native language—is stronglylinked to infants’ abilities to process words andsentences. This is important theoretically but isalso vital to the eventual use of these early pre-cursors to speech to diagnose children with de-velopmental disabilities that involve language.Furthermore, the fact that the earliest stagesof learning affect brain processing of both thesignals being learned (native patterns) and thesignals to which the infant is not exposed (non-native patterns) may play a role in our un-derstanding of the mechanisms underlying thecritical period, at least at the phonetic level,showing that learning itself, not merely time,affects one’s future ability to learn.

Whole-brain imaging now allows us to ex-amine multiple brain areas that are responsiveto speech, including those responsible for per-ception as well as production, revealing the pos-sible existence of a mirror system for speech.Research has begun to use these measures tounderstand how the bilingual brain maps twodistinct languages. Brain studies are beginningto show, in children with developmental disabil-ities, how particular disabilities affect the brainstructures underlying language learning. An-swers to the classic questions about the unique


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

human capacity to acquire language will con-tinue to be enriched by studies that utilize mod-

ern neuroscience tools to examine the infantbrain.

FUTURE DIRECTIONS

1. Are the brain structures activated in infants in response to language the same as thoseactivated in adults, and in both cases, are these brain systems speech specific?

2. Why do infants fail to learn language from television presentations—how does socialinteraction during language exposure affect the brain’s ability to learn?

3. Is the neural network connecting speech perception and speech production innate, andif so, is this network activated exclusively in response to language?

4. How is language mapped in the bilingual brain? Does experience with two or more lan-guages early in development affect the brain systems underlying social and/or cognitiveprocessing?

5. How do developmental disabilities such as autism, dyslexia, and specific language im-pairment affect the brain’s processing of speech?

6. Which causal mechanisms underlie the critical period for second language acquisition—why are adults, with their superior cognitive skills, unable to learn as well as younginfants? Can techniques be developed to help adults learn a second language?

DISCLOSURE STATEMENTThe authors are not aware of any biases that might be perceived as affecting the objectivity of thisreview.

ACKNOWLEDGMENTSThe authors are supported by a grant from the National Science Foundation Science of LearningCenter Program to the University of Washington LIFE Center, by grants to P.K. from the NationalInstitutes of Health (HD 37954, HD 34565, HD 55782).

LITERATURE CITEDAslin RN, Mehler J. 2005. Near-infrared spectroscopy for functional studies of brain activity in

human infants: promise, prospects, and challenges. J. Biomed. Opt. 10:011009Aslin RN, Saffran JR, Newport EL. 1998. Computation of conditional probability statistics by

8-month-old infants. Psychol. Sci. 9:321–24Bailey TM, Plunkett K. 2002. Phonological specificity in early words. Cogn. Dev. 17:1265–82Ballem KD, Plunkett K. 2005. Phonological specificity in children at 1;2. J. Child Lang. 32:159–73Baptista LF, Petrinovich L. 1986. Song development in the white-crowned sparrow: Social factors

and sex differences. Anim. Behav. 34:1359–71Best C, McRoberts GW. 2003. Infant perception of non-native consonant contrasts that adults

assimilate in different ways. Lang. Speech. 46:183–216Bongaerts T, Planken B, Schils E. 1995. Can late starters attain a native accent in a foreign lan-

guage: a test of the critical period hypothesis. In The Age Factor in Second Language Acquisition,ed. D Singleton, Z Lengyel, pp. 30–50. Clevedon, UK: Multiling. Matters


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Bortfeld H, Wruck E, Boas DA. 2007. Assessing infants’ cortical response to speech using near-infrared spectroscopy. Neuroimage. 34:407–15

Bortfeld HM, Morgan JL, Golinkoff RM, Rathbun K. 2005. Mommy and me: familiar names helplaunch babies into speech-stream segmentation. Psychol. Sci. 16:298–304

Bosch L, Sebastian-Galles N. 2003a. Language experience and the perception of a voicing contrastin fricatives: infant and adult data. Proc. Int. Congr. Phon. Sci., 15th, Barcelona, ed. MJ Sole, DRecasens, pp. 1987–90. Adelaide, Aust.: Casual Prod.

Bosch L, Sebastian-Galles N. 2003b. Simultaneous bilingualism and the perception of a language-specific vowel contrast in the first year of life. Lang. Speech. 46:217–43

Brainard MS, Knudsen EI. 1998. Sensitive periods for visual calibration of the auditory space mapin the barn owl optic tectum. J. Neurosci. 18:3929–42

Burnham D, Kitamura C, Vollmer-Conna U. 2002. What’s new, pussycat? On talking to babiesand animals. Science. 296:1435

Burns TC, Yoshida KA, Hill K, Werker JF. 2007. The development of phonetic representationin bilingual and monolingual infants. Appl. Psycholinguist. 28:455–74

Canseco-Gonzalez E. 2000. Using the recording of event-related brain potentials in the study ofsentence processing. In Language and the Brain: Representation and Processing, pp. 229–66. SanDiego: Academic

Cheour M, Ceponiene R, Lehtokoski A, Luuk A, Allik J, et al. 1998. Development of language-specific phoneme representations in the infant brain. Nat. Neurosci. 1:351–53

Cheour M, Imada T, Taulu S, Ahonen A, Salonen J, et al. 2004. Magnetoencephalography (MEG)is feasible for infant assessment of auditory discrimination. Exp. Neurol. 190:S44–51

Coffey-Corina S, Padden D, Kuhl PK, Dawson G. 2007. Electrophysiological processing of singlewords in toddlers and school-age children with Autism Spectrum Disorder. Presented at Annu.Meet. Cogn. Neurosci. Soc., New York

Conboy BT, Brooks R, Meltzoff AN, Kuhl PK. 2008. Joint engagement with language tutors pre-dicts brain and behavioral responses to second-language phonetic stimuli. Presented at ICIS, 16th,Vancouver

Conboy BT, Kuhl PK. 2007. ERP mismatch negativity effects in 11-month-old infants after exposureto Spanish. Presented at SRCD Bien Meet., Boston

Conboy BT, Mills DL. 2006. Two languages, one developing brain: event-related potentials towords in bilingual toddlers. Dev. Sci. 9:F1–12

Conboy BT, Rivera-Gaxiola M, Klarman L, Aksoylu E, Kuhl PK. 2005. Associations betweennative and nonnative speech sound discrimination and language development at the end ofthe first year. Proc. Suppl. Annu. Boston Univ. Conf. Lang. Dev., 29th, Boston http://www.bu.edu/linguistics/APPLIED/BUCLD/supp29.html

Cutler A, Norris D. 1988. The role of strong syllables in segmentation for lexical access. J. Exp.Psychol. Hum. Percept. Perform. 14:113–21

Dehaene-Lambertz G, Dehaene S, Hertz-Pannier L. 2002. Functional neuroimaging of speechperception in infants. Science. 298:2013–15

Dehaene-Lambertz G, Hertz-Pannier L, Dubois J, Meriaux S, Roche A, et al. 2006. Functionalorganization of perisylvian activation during presentation of sentences in preverbal infants.Proc. Natl. Acad. Sci. USA 103:14240–45

Eimas PD, Siqueland ER, Jusczyk P, Vigorito J. 1971. Speech perception in infants. Science.171:303–6

Fennell CT, Werker JF. 2003. Early word learners’ ability to access phonetic detail in well-knownwords. Lang. Speech. 3:245–64

Fernald A, Kuhl P. 1987. Acoustic determinants of infant preference for Motherese speech. InfantBehav. Dev. 10:279–93


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Fernald A, Perfors A, Marchman VA. 2006. Picking up speed in understanding: speech processingefficiency and vocabulary growth across the 2nd year. Dev. Psychol. 42:98–116

Fernald A, Simon T. 1984. Expanded intonation contours in mothers’ speech to newborns. Dev.Psychol. 20:104–13

Fowler CA. 1986. An event approach to the study of speech perception from a direct-realistperspective. J. Phon. 14:3–28

Friederici AD. 2002. Towards a neural basis of auditory sentence processing. Trends Cogn. Sci.6:78–84

Friederici AD. 2005. Neurophysiological markers of early language acquisition: from syllables tosentences. Trends Cogn. Sci. 9:481–88

Friederici AD, Fiebach CJ, Schlesewsky M, Bornkessel ID, von Cramon DY. 2006. Processinglinguistic complexity and grammaticality in the left frontal cortex. Cereb. Cortex. 16:1709–17

Friedrich M, Friederici AD. 2005. Lexical priming and semantic integration reflected in the event-related potential of 14-month-olds. NeuroReport. 16:653–56

Friedrich M, Friederici AD. 2006. Early N400 development and later language acquisition.Psychophysiology. 43:1–12

Gallese V. 2003. The manifold nature of interpersonal relations: the quest for a common mecha-nism. Philos. Trans. R. Soc. B 358:517–28

Ganger J, Brent MR. 2004. Reexamining the vocabulary spurt. Dev. Psychol. 40:621–32Gernsbacher MA, Kaschak MP. 2003. Neuroimaging studies of language production and com-

prehension. Annu. Rev. Psychol. 54:91–114Goldstein M, King A, West M. 2003. Social interaction shapes babbling: testing parallels between

birdsong and speech. Proc. Natl. Acad. Sci. USA 100:8030–35Grieser DL, Kuhl PK. 1988. Maternal speech to infants in a tonal language: support for universal

prosodic features in motherese. Dev. Psychol. 24:14–20Hahne A, Eckstein K, Friederici AD. 2004. Brain signatures of syntactic and semantic processes

during children’s language development. J. Cogn. Neurosci. 16:1302–18Hahne A, Jescheniak JD. 2001. What’s left if the jabberwock gets the semantics? An ERP investi-

gation into semantic and syntactic processes during auditory sentence comprehension. Cogn.Brain Res. 11:199–212

Halle P, de Boysson-Bardies B. 1994. Emergence of an early receptive lexicon: infants’ recognitionof words. Infant Behav. Dev. 17:119–29

Harris A. 2001. Processing semantic and grammatical information in auditory sentences: electro-physiological evidence from children and adults. Diss. Abstr. Int. Sect. B. Sci. Eng. 61:6729

Hauser MD, Weiss D, Marcus G. 2002. Rule learning by cotton-top tamarins. Cognition. 86:B15–22

Hohle B, Bijeljac-Babic R, Nazzi T, Herold B, Weissenborn J. 2008. The emergence of languagespecific prosodic preferences during the first half year of life: evidence from German andFrench infants. Cogn. Psychol. In press

Holcomb PJ, Coffey SA, Neville HJ. 1992. Visual and auditory sentence processing: a develop-mental analysis using event-related brain potentials. Dev. Neuropsychol. 8:203–41

Homae F, Watanabe H, Nakano T, Asakawa K, Taga G. 2006. The right hemisphere of sleepinginfant perceives sentential prosody. Neurosci. Res. 54:276–80

Imada T, Zhang Y, Cheour M, Taulu S, Ahonen A, et al. 2006. Infant speech perception activatesBroca’s area: a developmental magnetoencephalography study. NeuroReport 17:957–62

Immelmann K. 1969. Song development in the zebra finch and other estrildid finches. In BirdVocalizations, ed. R Hinde, pp. 61–74. London: Cambridge Univ. Press

Johnson EK, Jusczyk PW. 2001. Word segmentation by 8-month-olds: when speech cues countmore than statistics. J. Mem. Lang. 44:548–67


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Johnson J, Newport E. 1989. Critical period effects in sound language learning: the influence ofmaturation state on the acquisition of English as a second language. Cogn. Psychol. 21:60–99

Jusczyk PW. 1997. The Discovery of Spoken Language. Cambridge, MA: MIT PressJusczyk PW, Aslin R. 1995. Infants’ detection of the sound patterns of words in fluent speech.

Cogn. Psychol. 29:1–23Jusczyk PW, Hohne EA. 1997. Infants’ memory for spoken words. Science. 277:1984–86Jusczyk PW, Rosner BS, Cutting JE, Foard CF, Smith LB. 1977. Categorical perception of

nonspeech sounds by 2-month-old infants. Percep. Psychophys. 21:50–54Klarman L, Krentz U, Brinkley B, Corina D, Kuhl PK. 2007. Deaf and hearing infants preference for

American Sign Language. Poster presented at Annu. Meet. Am. Psychol. Assoc., 115th, SanFrancisco

Krentz UC, Corina DP. 2008. Preference for language in early infancy: the human language biasis not speech specific. Dev. Sci. 11:1–9

Kuhl PK. 2000. A new view of language acquisition. Proc. Natl. Acad. Sci. USA 97:11850–57Kuhl PK. 2004. Early language acquisition: cracking the speech code. Nat. Rev. Neurosci. 5:831–43Kuhl PK. 2007. Is speech learning ‘gated’ by the social brain? Dev. Sci. 10:110–20Kuhl PK, Andruski JE, Chistovich IA, Chistovich LA, Kozhevnikova EV, et al. 1997. Cross-

language analysis of phonetic units in language addressed to infants. Science. 277:684–86Kuhl PK, Coffey-Corina S, Padden D, Dawson G. 2005a. Links between social and linguistic

processing of speech in preschool children with autism: Behavioral and electrophysiologicalevidence. Dev. Sci. 8:1–12

Kuhl PK, Conboy BT, Coffey-Corina S, Padden D, Nelson T, et al. 2005b. Early speech percep-tion and later language development: implications for the “critical period.” Lang. Learn. Dev.1:237–64

Kuhl PK, Conboy BT, Padden D, Rivera-Gaxiola M, Nelson T. 2008. Phonetic learning asa pathway to language: new data and native language magnet theory expanded (NLM-e).Philos. Trans. R. Soc. B. 363:979–1000

Kuhl PK, Damasio A. 2008. Language. In Principles of Neural Science, ed. ER Kandel, JH Schwartz,TM Jessell, S Siegelbaum, J Hudspeth. New York: McGraw Hill. 5th ed. In press

Kuhl PK, Meltzoff AN. 1982. The bimodal perception of speech in infancy. Science. 218:1138–41Kuhl PK, Meltzoff AN. 1996. Infant vocalizations in response to speech: vocal imitation and

developmental change. J. Acoust. Soc. Am. 100:2425–38Kuhl PK, Miller JD. 1975. Speech perception by the chinchilla: voiced-voiceless distinction in

alveolar plosive consonants. Science. 190:69–72Kuhl PK, Stevens E, Hayashi A, Deguchi T, Kiritani S, et al. 2006. Infants show a facilitation

effect for native language phonetic perception between 6 and 12 months. Dev. Sci. 9:F13–21Kuhl PK, Tsao F-M, Liu H-M. 2003. Foreign-language experience in infancy: effects of short-term

exposure and social interaction on phonetic learning. Proc. Natl. Acad. Sci. USA 100:9096–101Kuhl PK, Williams KA, Lacerda F, Stevens KN, Lindblom B. 1992. Linguistic experience alters

phonetic perception in infants by 6 months of age. Science. 255:606–8Kujala A, Alho K, Service E, Ilmoniemi RJ, Connolly JF. 2004. Activation in the anterior left

auditory cortex associated with phonological analysis of speech input: localization of thephonological mismatch negativity response with MEG. Cogn. Brain Res. 21:106–13

Kutas M. 1997. Views on how the electrical activity that the brain generates reflects the functionsof different language structures. Psychophysiology. 34:383–98

Lewkowicz DJ, Ghazanfar AA. 2006. The decline of cross-species intersensory perception inhuman infants. Proc. Natl. Acad. Sci. USA 103:6771–74

Liberman AM, Cooper FS, Shankweiler DP, Studdert-Kennedy M. 1967. Perception of the speechcode. Psychol. Rev. 74:431–61


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Liberman AM, Mattingly IG. 1985. The motor theory of speech perception revised. Cognition.21:1–36

Liu H-M, Kuhl PK, Tsao F-M. 2003. An association between mothers’ speech clarity and infants’speech discrimination skills. Dev. Sci. 6:F1–10

Liu H-M, Tsao F-M, Kuhl PK. 2007. Acoustic analysis of lexical tone in Mandarin infant-directedspeech. Dev. Psychol. 43:912–17

Mandel DR, Jusczyk PW, Pisoni D. 1995. Infants’ recognition of the sound patterns of their ownnames. Psychol. Sci. 6:314–17

Mayberry RI, Lock E. 2003. Age constraints on first versus second language acquisition: evidencefor linguistic plasticity and epigenesis. Brain Lang. 87:369–84

Maye J, Werker JF, Gerken L. 2002. Infant sensitivity to distributional information can affectphonetic discrimination. Cognition. 82:B101–11

Meltzoff AN, Decety J. 2003. What imitation tells us about social cognition: a rapprochement be-tween developmental psychology and cognitive neuroscience. Philos. Trans R. Soc. B. 358:491–500

Meltzoff AN, Moore MK. 1997. Explaining facial imitation: a theoretical model. Early Dev. Parent.6:179–92

Mills DL, Coffey-Corina SA, Neville HJ. 1993. Language acquisition and cerebral specializationin 20-month-old infants. J. Cogn. Neurosci. 5:317–34

Mills DL, Coffey-Corina S, Neville HJ. 1997. Language comprehension and cerebral specializa-tion from 13–20 months. Dev. Neuropsychol. 13:233–37

Mills DL, Plunkett K, Prat C, Schafer G. 2005. Watching the infant brain learn words: effects ofvocabulary size and experience. Cogn. Dev. 20:19–31

Mills DL, Prat C, Zangl R, Stager CL, Neville HJ, et al. 2004. Language experience and theorganization of brain activity to phonetically similar words: ERP evidence from 14- and20-month-olds. J. Cogn. Neurosci. 16:1452–64

Molfese DL. 1990. Auditory evoked responses recorded from 16-month-old human infants towords they did and did not know. Brain Lang. 38:345–63

Molfese DL, Morse PA, Peters CJ. 1990. Auditory evoked responses to names for different objects:cross-modal processing as a basis for infant language acquisition. Dev. Psychol. 26:780–95

Molfese DL, Wetzel WF, Gill LA. 1993. Known versus unknown word discriminations in 12-month-old human infants: electrophysiological correlates. Dev. Neuropsychol. 9:241–58

Munte TF, Matzke M, Johanes S. 1997. Brain activity associated with syntactic incongruencies inwords and psuedowords. J. Cogn. Neurosci. 9:318–29

Naatanen R, Lehtokoski A, Lennes M, Cheour M, Huotilainen M, et al. 1997. Language-specificphoneme representations revealed by electric and magnetic brain responses. Nature. 385:432–34

Nazzi TL, Iakimova G, Bertoncini J, Fredonie S, Alcantara C. 2006. Early segmentation of fluentspeech by infants acquiring French: emerging evidence for crosslinguistic differences. J. Mem.Lang. 54:283–99

Neville HJ, Coffey SA, Holcomb PJ, Tallal P. 1993. The neurobiology of sensory and languageprocessing in language-impaired children. Cogn. Neurosci. 5:235–53

Neville HJ, Coffey SA, Lawson DS, Fischer A, Emmorey K, et al. 1997. Neural systems mediatingAmerican Sign Language: effects of sensory experience and age of acquisition. Brain Lang.57:285–308

Neville HJ, Mills DL, Lawson DS. 1992. Fractionating language: different neural subsystems withdifferent sensitive periods. Cereb. Cortex 2:244–58


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Newman R, Ratner NB, Jusczyk AM, Jusczyk PW, Dow KA. 2006. Infants’ early ability to segmentthe conversational speech signal predicts later language development: a retrospective analysis.Dev. Psychol. 42:643–55

Newport E. 1990. Maturational constraints on language learning. Cogn. Sci. 14:11–28Newport EL, Aslin RN. 2004a. Learning at a distance I. Statistical learning of nonadjacent de-

pendencies. Cogn. Psychol. 48:127–62Newport EL, Aslin RN. 2004b. Learning at a distance II. Statistical learning of nonadjacent

dependencies in a nonhuman primate. Cogn. Psychol. 49:85–117Noble KG, McCandliss BD, Farah MJ. 2007. Socioeconomic gradients predict individual differ-

ences in neurocognitive abilities. Dev. Sci. 10:464–80Noble KG, Wolmetz ME, Ochs LG, Farah MJ, McCandliss BD. 2006. Brain-behavior relation-

ships in reading acquisition are modulated by socioeconomic factors. Dev. Sci. 9:642–54Oberecker R, Friederici AD. 2006. Syntactic event-related potential components in 24-month-

olds’ sentence comprehension. NeuroReport 17:1017–21Oberecker R, Friedrich M, Friederici AD. 2005. Neural correlates of syntactic processing in

two-year-olds. J. Cogn. Neurosci. 17:1667–78Pena M, Bonatti L, Nespor M, Mehler J. 2002. Signal-driven computations in speech processing.

Science 298:604–7Pulvermuller F. 2005. The neuroscience of language: on brain circuits of words and serial order.

Cambridge, UK: Med. Res. Counc., Cambridge Univ. PressRaudenbush SW, Bryk AS, Cheong YF, Congdon R. 2005. HLM-6: Hierarchical Linear and Non-

linear Modeling. Lincolnwood, IL: Sci. Softw. Int.Raizada RD, Richards TL, Meltzoff AN, Kuhl PK. 2008. Socioeconomic status predicts hemi-

spheric specialization of the left inferior frontal gyrus in young children. NeuroImage.40:1392–401

Rivera-Gaxiola M, Klarman L, Garcia-Sierra A, Kuhl PK. 2005a. Neural patterns to speech andvocabulary growth in American infants. NeuroReport 16:495–98

Rivera-Gaxiola M, Silva-Pereyra J, Kuhl PK. 2005b. Brain potentials to native and non-nativespeech contrasts in 7- and 11-month-old American infants. Dev. Sci. 8:162–72

Rizzolatti G. 2005. The mirror neuron system and its function in humans. Anat. Embryol. 210:419–21

Rizzolatti G, Craighero L. 2004. The mirror neuron system. Annu. Rev. Neurosci. 27:169–92Rizzolatti G, Fadiga L, Fogassi L, Gallese V. 2002. From mirror neurons to imitation, facts, and

speculations. In The Imitative Mind: Development, Evolution, and Brain Bases, ed. AN Meltzoff,W Prinz, pp. 247–66. Cambridge, UK: Cambridge Univ. Press

Rizzolatti G, Fadiga L, Gallese V, Fogassi L. 1996. Premotor cortex and the recognition of motoractions. Cogn. Brain Res. 3:131–41

Saffran JR. 2003. Statistical language learning: mechanisms and constraints. Curr. Dir. Psychol. Sci.12:110–14

Saffran JR, Aslin RN, Newport EL. 1996. Statistical learning by 8-month old infants. Science.274:1926–28

Saffran JR, Johnson EK, Aslin RN, Newport EL. 1999. Statistical learning of tone sequences byhuman infants and adults. Cognition. 70:27–52

Silva-Pereyra J, Conboy BT, Klarman L, Kuhl PK. 2007. Grammatical processing without se-mantics? An event-related brain potential study of preschoolers using jabberwocky sentences.J. Cogn. Neurosci. 19:1050–65

Silva-Pereyra J, Klarman L, Lin Jo-Fu L, Kuhl PK. 2005a. Sentence processing in 30-month-oldchildren: an ERP study. NeuroReport. 16:645–48


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

ANRV346-NE31-21 ARI 14 May 2008 12:46

Silva-Pereyra J, Rivera-Gaxiola M, Kuhl PK. 2005b. An event-related brain potential study ofsentence comprehension in preschoolers: semantic and morphosyntatic processing. Cogn.Brain Res. 23:247–85

Silven M, Kouvo A, Haapakoski M, Lahteenmaki V, Voeten M, et al. 2006. Early speech perceptionand vocabulary growth in bilingual infants exposed to Finnish and Russian. Poster presented atLang. Acquis. Biling. Conf., Toronto

Stager CL, Werker JF. 1997. Infants listen for more phonetic detail in speech perception than inword-learning tasks. Nature. 388:381–82

Streeter LA. 1976. Language perception of 2-month-old infants shows effects of both innatemechanisms and experience. Nature. 259:39–41

Sundara M, Polka L, Genesee F. 2006. Language-experience facilitates discrimination of /d- / inmonolingual and bilingual acquisition of English. Cognition 100:369–88

Swingley D. 2005. 11-month-olds’ knowledge of how familiar words sound. Dev. Sci. 8:432–43Swingley D, Aslin RN. 2000. Spoken word recognition and lexical representation in very young

children. Cognition. 76:147–66Swingley D, Aslin RN. 2002. Lexical neighborhoods and the word-form representations of 14-

month-olds. Psychol. Sci. 13:480–84Swingley D, Aslin RN. 2007. Lexical competition in young children’s word learning. Cogn. Psychol.

54:99–132Taga G, Asakawa K. 2007. Selectivity and localization of cortical response to auditory and visual

stimulation in awake infants aged 2 to 4 months. Neuroimage 36:1246–52Thierry G, Vihman M, Roberts M. 2003. Familiar words capture the attention of 11-month-olds

in less than 250 ms. NeuroReport. 14:2307–10Thiessen ED, Saffran JR. 2003. When cues collide: use of stress and statistical cues to word

boundaries by 7- to 9-month-old infants. Dev. Psychol. 39:706–16Tincoff R, Jusczyk PW. 1999. Some beginnings of word comprehension in 6-month-olds. Psychol.

Sci. 10:172–75Tsao F-M, Liu H-M, Kuhl PK. 2004. Speech perception in infancy predicts language development

in the second year of life: a longitudinal study. Child Dev. 75:1067–84Vihman MM, Nakai S, DePaolis RA, Halle P. 2004. The role of accentual pattern in early lexical

representation. J. Mem. Lang. 50:336–53Vouloumanos A, Werker JF. 2004. Tuned to the signal: the privileged status of speech for young

infants. Dev. Sci. 7:270–76Weber C, Hahne A, Friedrich M, Friederici AD. 2004. Discrimination of word stress in early

infant perception: electrophysiological evidence. Cogn. Brain Res. 18:149–61Weber-Fox C, Neville HJ. 1996. Maturational constraints on functional specializations for lan-

guage processing: ERP and behavioral evidence in bilingual speakers. J. Cogn. Neurosci. 8:231–56

Weikum WM, Vouloumanos A, Navarra J, Soto-Faraco S, Sebastian-Galles N, et al. 2007. Visuallanguage discrimination in infancy. Science 316:1159

Werker JF, Curtin S. 2005. PRIMIR: a developmental framework of infant speech processing.Lang. Learn. Dev. 1:197–234

Werker JF, Fennell CT, Corcoran KM, Stager CL. 2002. Infants’ ability to learn phoneticallysimilar words: effects of age and vocabulary size. Infancy. 3:1–30

Werker JF, Tees RC. 1984. Cross-language speech perception: evidence for perceptual reorgani-zation during the first year of life. Infant Behav. Dev. 7:49–63

Zhang Y, Kuhl PK, Imada T, Kotani M, Tohkura Y. 2005. Effects of language experience: neuralcommitment to language-specific auditory patterns. NeuroImage 26:703–20


Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

AR346-FM ARI 20 May 2008 15:1

Annual Review ofNeuroscience

Volume 31, 2008Contents

Cerebellum-Like Structures and Their Implications for CerebellarFunctionCurtis C. Bell, Victor Han, and Nathaniel B. Sawtell ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 1

Spike Timing–Dependent Plasticity: A Hebbian Learning RuleNatalia Caporale and Yang Dan ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !25

Balancing Structure and Function at Hippocampal Dendritic SpinesJennifer N. Bourne and Kristen M. Harris ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !47

Place Cells, Grid Cells, and the Brain’s Spatial Representation SystemEdvard I. Moser, Emilio Kropff, and May-Britt Moser ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !69

Mitochondrial Disorders in the Nervous SystemSalvatore DiMauro and Eric A. Schon ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !91

Vestibular System: The Many Facets of a Multimodal SenseDora E. Angelaki and Kathleen E. Cullen ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 125

Role of Axonal Transport in Neurodegenerative DiseasesKurt J. De Vos, Andrew J. Grierson, Steven Ackerley, and Christopher C.J. Miller ! ! ! 151

Active and Passive Immunotherapy for Neurodegenerative DisordersDavid L. Brody and David M. Holtzman ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 175

Descending Pathways in Motor ControlRoger N. Lemon ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 195

Task Set and Prefrontal CortexKatsuyuki Sakai ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 219

Multiple Sclerosis: An Immune or Neurodegenerative Disorder?Bruce D. Trapp and Klaus-Armin Nave ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 247

Multifunctional Pattern-Generating CircuitsK.L. Briggman and W.B. Kristan, Jr. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 271

Retinal Axon Growth at the Optic Chiasm: To Cross or Not to CrossTimothy J. Petros, Alexandra Rebsam, and Carol A. Mason ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 295

v

Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

AR346-FM ARI 20 May 2008 15:1

Brain Circuits for the Internal Monitoring of MovementsMarc A. Sommer and Robert H. Wurtz ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 317

Wnt Signaling in Neural Circuit AssemblyPatricia C. Salinas and Yimin Zou ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 339

Habits, Rituals, and the Evaluative BrainAnn M. Graybiel ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 359

Mechanisms of Self-Motion PerceptionKenneth H. Britten ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 389

Mechanisms of Face PerceptionDoris Y. Tsao and Margaret S. Livingstone ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 411

The Prion’s Elusive Reason for BeingAdriano Aguzzi, Frank Baumann, and Juliane Bremer ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 439

Mechanisms Underlying Development of Visual Maps andReceptive FieldsAndrew D. Huberman, Marla B. Feller, and Barbara Chapman ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 479

Neural Substrates of Language AcquisitionPatricia K. Kuhl and Maritza Rivera-Gaxiola ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 511

Axon-Glial Signaling and the Glial Support of Axon FunctionKlaus-Armin Nave and Bruce D. Trapp ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 535

Signaling Mechanisms Linking Neuronal Activity to Gene Expressionand Plasticity of the Nervous SystemSteven W. Flavell and Michael E. Greenberg ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 563

Indexes

Cumulative Index of Contributing Authors, Volumes 22–31 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 591

Cumulative Index of Chapter Titles, Volumes 22–31 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 595

Errata

An online log of corrections to Annual Review of Neuroscience articles may be found athttp://neuro.annualreviews.org/

vi Contents

Ann

u. R

ev. N

euro

sci.

2008

.31:

511-

534.

Dow

nloa

ded

from

arjo

urna

ls.an

nual

revi

ews.o

rgby

UN

IVER

SITY

OF

WA

SHIN

GTO

N -

HEA

LTH

SCI

ENCE

S LI

BRA

RIES

on

06/3

0/08

. For

per

sona

l use

onl

y.

Neural Substrates of Language Acquisition · of Language Acquisition ... effortlessly,...

Documents

Transcript of Neural Substrates of Language Acquisition · of Language Acquisition ... effortlessly,...